Not sure the first unit test is right in week 4 lab

Course Q&A Deep Learning Specialization Sequence Models
,
2

It worked for me. I added some print statements in the body of get_angles to show what is going on and here’s what I see:

pos [[0]
 [1]
 [2]
 [3]]
k [[ 0  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15]]
angles [[0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00
  0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00
  0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00
  0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00]
 [1.00000000e+00 1.00000000e+00 3.16227766e-01 3.16227766e-01
  1.00000000e-01 1.00000000e-01 3.16227766e-02 3.16227766e-02
  1.00000000e-02 1.00000000e-02 3.16227766e-03 3.16227766e-03
  1.00000000e-03 1.00000000e-03 3.16227766e-04 3.16227766e-04]
 [2.00000000e+00 2.00000000e+00 6.32455532e-01 6.32455532e-01
  2.00000000e-01 2.00000000e-01 6.32455532e-02 6.32455532e-02
  2.00000000e-02 2.00000000e-02 6.32455532e-03 6.32455532e-03
  2.00000000e-03 2.00000000e-03 6.32455532e-04 6.32455532e-04]
 [3.00000000e+00 3.00000000e+00 9.48683298e-01 9.48683298e-01
  3.00000000e-01 3.00000000e-01 9.48683298e-02 9.48683298e-02
  3.00000000e-02 3.00000000e-02 9.48683298e-03 9.48683298e-03
  3.00000000e-03 3.00000000e-03 9.48683298e-04 9.48683298e-04]]
All tests passed
pos [[0]
 [1]
 [2]
 [3]]
k [[0 1 2 3 4 5 6 7]]
angles [[0.e+00 0.e+00 0.e+00 0.e+00 0.e+00 0.e+00 0.e+00 0.e+00]
 [1.e+00 1.e+00 1.e-01 1.e-01 1.e-02 1.e-02 1.e-03 1.e-03]
 [2.e+00 2.e+00 2.e-01 2.e-01 2.e-02 2.e-02 2.e-03 2.e-03]
 [3.e+00 3.e+00 3.e-01 3.e-01 3.e-02 3.e-02 3.e-03 3.e-03]]

So you can see that (at least this time :smile:) the docstring of the function is actually correct: pos is a column vector and k is a row vector. So you don’t need to do any manipulation of the shapes or orientations: just write the code as if they were scalars and it all “just works” through the magic of broadcasting.

To see a demonstration of what is happening there, watch this:

A = np.arange(4)[:, np.newaxis]
B = 2 * np.ones((1,6))
print(f"A.shape {A.shape}")
print(f"A {A}")
print(f"B.shape {B.shape}")
print(f"B {B}")
quotient = A / B
print(f"quotient.shape {quotient.shape}")
print(f"quotient {quotient}")
A.shape (4, 1)
A [[0]
 [1]
 [2]
 [3]]
B.shape (1, 6)
B [[2. 2. 2. 2. 2. 2.]]
quotient.shape (4, 6)
quotient [[0.  0.  0.  0.  0.  0. ]
 [0.5 0.5 0.5 0.5 0.5 0.5]
 [1.  1.  1.  1.  1.  1. ]
 [1.5 1.5 1.5 1.5 1.5 1.5]]

The “/” operation is “elementwise”, so broadcasting just expands both operands to get a compatible shape.

Also note that we aren’t doing Fourier Transforms yet. That doesn’t come until the next function. The reason the first row sums to zero is that it’s all zeros, right? :nerd_face:

Grader output: seems fishy